[molpro-user] parallel calculations for molpro 2010 version

[lydia]

Dear Molpro users,
I recently have some strange calculations for parallel jobs using the
molpro 2010 version?
(1) It happens frequently that the jobs are hanging there without any
error message. It will be ok if I submit it again.
(2)For a single-point CASSCF calculation, it converged with
one-processor in 4 iterations.The 4,6,8 processers were also tried but
get non-converged wavefunction in default 20 iterations.
(3)The same frequencies calculation job, one processer, but submiited
to different node, one is converged, the other one is not.
(4) The same single point CASSCF calculation job, one processer, but
submiited to different node, they used different iterations to
converge and get a little different energies.
  The output for CASSCF wavefunction part are shown here.
   One case:
   ITER. MIC  NCI  NEG     ENERGY(VAR)     ENERGY(PROJ)   ENERGY
CHANGE     GRAD(0)  GRAD(ORB)   GRAD(CI)     STEP       TIME
   1  201   24    0   -8636.43352229   -8636.64027685   -0.20675456
0.11489394 0.01676296 0.46676481  0.22D+01   1176.20
   2   90   52    0   -8636.64820415   -8636.67151650   -0.02331234
0.09028047 0.00000831 0.00446933  0.82D+00   2685.35
   3  100   55    0   -8636.67159127   -8636.67168352   -0.00009225
0.00703551 0.00000081 0.00395209  0.48D-01   4333.36
   4  110   52    0   -8636.67169167   -8636.67171455   -0.00002288
0.00364815 0.00000019 0.00298435  0.24D-01   5979.92
   5  110   51    0   -8636.67172002   -8636.67173417   -0.00001415
0.00251066 0.00000005 0.00082504  0.23D-01   7525.23
   6  110   26    0   -8636.67173446   -8636.67173446    0.00000000
0.00066144 0.00000269 0.00001017  0.00D+00   8690.29
** WVFN ****  CONVERGENCE REACHED, FINAL GRADIENT:  0.66D-03
TOTAL ENERGIES                     -8636.68736874 -8636.65610019

and the other case
ITER. MIC  NCI  NEG     ENERGY(VAR)     ENERGY(PROJ)   ENERGY CHANGE
  GRAD(0)  GRAD(ORB)   GRAD(CI)     STEP       TIME
   1  201   22    0   -8636.43352227   -8636.58589577   -0.15237350
0.11489394 0.01528358 0.45149325  0.22D+01    515.57
   2   60   46    0   -8636.59016316   -8636.66885304   -0.07868988
0.05439846 0.00039155 0.07838618  0.13D+01   1309.59
   3   60   41    0   -8636.67094504   -8636.67153988   -0.00059483
0.02521905 0.00000205 0.02061308  0.90D-01   2081.21
   4   60   41    0   -8636.67154760   -8636.67153492    0.00001269
0.00371838 0.00000182 0.00744618  0.38D-02   2898.67
   5   60   40    0   -8636.67154090   -8636.67153105    0.00000985
0.00368663 0.00000237 0.01345678  0.31D-02   3652.02
   6   60   40    0   -8636.67153660   -8636.67152490    0.00001170
0.00359435 0.00000230 0.00439362  0.43D-02   4410.63
   7   60   40    0   -8636.67152997   -8636.67151905    0.00001093
0.00346552 0.00000212 0.01311886  0.41D-02   5164.94
   8   60   35    0   -8636.67152380   -8636.67150829    0.00001551
0.00335199 0.00000164 0.00970629  0.79D-02   5899.73
   9   60   36    0   -8636.67151223   -8636.67150150    0.00001073
0.00306869 0.00000160 0.00751871  0.53D-02   6711.25
  10   60   40    0   -8636.67150494   -8636.67149573    0.00000921
0.00287847 0.00000151 0.00529872  0.49D-02   7531.90
  11   60   40    0   -8636.67149868   -8636.67149089    0.00000779
0.00267814 0.00000136 0.00620809  0.44D-02   8310.69
  12   60   40    0   -8636.67149345   -8636.67148769    0.00000576
0.00249146 0.00000123 0.00598112  0.31D-02   9086.39
  13   60   40    0   -8636.67148998   -8636.67148650    0.00000348
0.00236137 0.00000119 0.00482774  0.12D-02   9857.55
  14   59   39    0   -8636.67148871   -8636.67149269   -0.00000398
0.00231938 0.00000130 0.01149000  0.60D-02  10682.80
  15   60   40    0   -8636.67149549   -8636.67149078    0.00000471
0.00258364 0.00000133 0.00536661  0.19D-02  11525.95
  16   60   40    0   -8636.67149336   -8636.67148984    0.00000352
0.00250317 0.00000129 0.00484937  0.97D-03  12357.74
 ** WVFN ****  CONVERGENCE REACHED, FINAL GRADIENT:  0.25D-02
TOTAL ENERGIES                     -8636.68809932 -8636.65488542

The situation above are serious and confused me a lot. They are
strange. Do you think this problem come from the the molpro codes?
I ever doubt that the sauce codes were not compiled and built
correctly in our clusters, but the worker in the computer center of my
university said that all the test jobs are passed successfully. So I
put my questions here and beg the experts give me some suggestions.
Thanks in advance.

All the best,
lydia